John Collison on the path dependence of startups

October 30, 2025

“There tends to be a lot of debate around the Lean Startup Methodology where you should have a product, show it to some customers, and rapidly iterate based on their usage and feedback versus the vision-driven product of, ‘I know what this is and I know where it should go.’ And to me, that’s not really a tradeoff at all. You need some small set of things — the vision — that you’re not willing to compromise on, and then if the market is telling you you’re wrong, you really need to listen to that data.”

In Stripe’s case, John explains, “We were really focused on building a developer-friendly tool as our path to market. If that wasn’t a valid way to build this company, I’m not sure we would’ve been able to pivot out of that. But in every other area we explored the market, we just kept running into feedback from customers that ended up being really critical to allowing the product to succeed.”

John continues:

“Startups end up having a really powerful path dependence where it’s not just a question of arriving at the final answer . . . It’s a question of building a set of assumptions, following those, and continually revisiting them. You’re not just trying to arrive at the final form of your product. You’re trying to get all the intermediate forms and the chemical reaction that gets you there.”

He gives Microsoft as an example:

“They started out in the hobbyist market building BASIC compilers. They couldn’t have stayed there and still become the company they are today building compilers. They had to move upstream, and they’re now basically an enterprise company. But at the same time, they couldn’t have started as an enterprise company because nobody would’ve bought from these two random college kids. There’s this path-dependence to Microsoft’s growth, where they had to become each one of their stages to allow them to jump to the next level.”

Uber is another example:

“If they had just started out with their Uber X model today doing ride sharing, they probably would’ve been shut down pretty quickly. To succeed they needed to start with the limo business, get to a scale where it was a well-known and well-liked service, and then they could start doing ride sharing in some cities and prove that it was a popular model.”